Speaker-independent features extracted by a neural network.

Y. Kato, Masahide Sugiyama

Browse the full ICASSP paper archive.